Robust and high-throughput lipidomic quantitation of human blood samples using flow injection analysis with tandem mass spectrometry for clinical use

Cohort-based strategies as an in-house tool to evaluate and improve phenotyping robustness of LC-MS/MS lipidomics platforms

In recent years, instrumental improvements have enabled the spread of mass spectrometry-based lipidomics platforms in biomedical research. In mass spectrometry, the reliability of generated data varies for each compound, contingent on, among other factors, the availability of labeled internal standards. It is challenging to evaluate the data for lipids without specific labeled internal standards, especially when dozens to hundreds of lipids are measured simultaneously. Thus, evaluation of the performance of these platforms at the individual lipid level in interlaboratory studies is generally not feasible in a time-effective manner. Herein, using a focused subset of sphingolipids, we present an in-house validation methodology for individual lipid reliability assessment, tailored to the statistical analysis to be applied. Moreover, this approach enables the evaluation of various methodological aspects, including discerning coelutions sharing identical selected reaction monitoring transitions, pinpointing optimal labeled internal standards and their concentrations, and evaluating different extraction techniques. While the full validation according to analytical guidelines for all lipids included in a lipidomics method is currently not possible, this process shows areas to focus on for subsequent method development iterations as well as the robustness of data generated across diverse methodologies.

Improved lipid analysis using a 2D-LC-MS system with a novel injection procedure

The aim of this study was to improve analysis of nonpolar lipidomics sample extracts using reversed phase (RP) chromatography. A 4/3/3 (v/v/v) mixture of methanol/methyl tert-butyl ether/chloroform (MeOH/MTBE/CHCl3, MMC) was chosen for sample extraction solvent based on its proven extraction capability for several lipid classes. To avoid carry over, loss of analytes and peak distortion the loops and all capillaries of the presented LC system were flushed and filled up with methanol until the analytical column. The choice of methanol was due to its weak elution strength and being infinitely miscible with MMC and several other nonpolar solvents. This allowed injection of a 100 μl sample that was 20 μl nonpolar extraction solvent diluted fivefold with methanol. All lipids of 25 lipid classes were transferred quantitatively to the column head where the online dilution of methanol was carried out with aqueous eluent for focusing the lipid analytes. The weak elution strength of methanol prevented peak distortions. The consecutive reversed phase elution resulted in remarkably narrow peaks (full width at half maximum was 0.07-0.08 min typically) and enhanced sensitivity (limit of detection usually in sub nM region) because of increased sample injection volume and narrow peaks. Calibration and quality control samples made by diluting commercial lipid standards 200-50000 times confirmed the applicability of this approach both for targeted lipid quantification and for untargeted quantitative comparison of lipids from different sources.

Sub-5-min RP-UHPLC-TIMS for high-throughput untargeted lipidomics and its application to multiple matrices

Untargeted lipidomics, with its ability to take a snapshot of the lipidome landscape, is an important tool to highlight lipid changes in pathology or drug treatment models. One of the shortcomings of most untargeted lipidomics based on UHPLC-HRMS is the low throughput, which is not compatible with large-scale screening. In this contribution, we evaluate the application of a sub-5-min high-throughput four-dimensional trapped ion mobility mass spectrometry (HT-4D-TIMS) platform for the fast profiling of multiple complex biological matrices. Human AC-16 cells and mouse brain, liver, sclera, and feces were used as samples. By using a fast 4-min RP gradient, the implementation of TIMS allows us to differentiate coeluting isomeric and isobaric lipids, with correct precursor ion isolation, avoiding co-fragmentation and chimeric MS/MS spectra. Globally, the HT-4D-TIMS allowed us to annotate 1910 different lipid species, 1308 at the molecular level and 602 at the sum composition level, covering 58 lipid subclasses, together with quantitation capability covering more than three orders of magnitude. Notably, TIMS values were highly comparable with respect to longer LC gradients (CV% = 0.39%). These results highlight how HT-4D-TIMS-based untargeted lipidomics possess high coverage and accuracy, halving the analysis time with respect to conventional UHPLC methods, and can be used for fast and accurate untargeted analysis of complex matrices to rapidly evaluate changes of lipid metabolism in disease models or drug discovery campaigns.

HILIC/MS quantitation of low-abundant phospholipids and sphingolipids in human plasma and serum: Dysregulation in pancreatic cancer

A new hydrophilic interaction liquid chromatography - mass spectrometry method is developed for low-abundant phospholipids and sphingolipids in human plasma and serum. The optimized method involves the Cogent Silica type C hydride column, the simple sample preparation by protein precipitation, and the removal of highly abundant lipid classes using the postcolumn valve directed to waste during two elution windows. The method allows a highly confident and sensitive identification of low-abundant lipid classes in human plasma (246 lipid species from 24 lipid subclasses) based on mass accuracy and retention dependencies in both polarity modes. The method is validated for quantitation using two internal standards (if available) for each lipid class and applied to human plasma and serum samples obtained from patients with pancreatic ductal adenocarcinoma (PDAC), healthy controls, and NIST SRM 1950. Multivariate data analysis followed by various statistical projection methods is used to determine the most dysregulated lipids. Significant downregulation is observed for lysophospholipids with fatty acyl composition 16:0, 18:0, 18:1, and 18:2. Distinct trends are observed for phosphatidylethanolamines (PE) in relation to the bonding type of fatty acyls, where most PE with acyl bonds are upregulated, while ether/plasmenyl PE are downregulated. For the sphingolipid category, sphingolipids with very long N-acyl chains are downregulated, while sphingolipids with shorter N-acyl chains were upregulated in PDAC. These changes are consistently observed for various classes of sphingolipids, ranging from ceramides to glycosphingolipids, indicating a possible metabolic disorder in ceramide biosynthesis caused by PDAC.

Selective ionization of oxidized lipid species using different solvent additives in flow injection mass spectrometry

Lipid oxidation in food products is a crucial problem that causes undesirable changes in the food's flavor, texture, and nutritional value. It should be carefully monitored as it can lead to the formation of potentially toxic compounds and in that way reduce the shelf life of the product. Liquid chromatography coupled to mass spectrometry is a powerful tool to monitor the formation of oxidized lipids. However, the presence of lipid species in both their non-oxidized and oxidized forms at distinctly different concentrations can hinder the detection and identification of the less abundant oxidized species, due to coelution. In this study, a flow injection mass spectrometry approach was used to selectively ionize oxidized triacylglycerols versus their non-oxidized precursors. Three mobile phase additives were investigated (ammonium formate, sodium acetate, and sodium iodide) at three different concentrations, and ion source settings (i.e., sheath gas temperature, capillary voltage, and nozzle voltage) were optimized. A fractional factorial design was conducted to examine not only the direct effect of the operating parameters on the selectivity of ionization for the oxidized lipid species, but also to assess their combined effect. Overall, selective ionization of oxidized versus non-oxidized lipid species was favored by the use of sodium-containing solvent additives. The application of specific ion source settings resulted in an increased ionization selectivity, with sheath gas temperature and capillary voltage having the most significant influence. A selectivity factor as high as 120 could be reached by combining 0.1 mg/mL sodium-containing additives, with 250 °C sheath gas temperature and 5000 V capillary voltage. These findings will contribute to future studies on fast detection and relative quantification of low abundant oxidized triacylglycerols and their possible impact on human health.

Total analysis system for the determination of uremic toxins in human plasma based on bead injection solid phase extraction hyphenated to mass spectrometry

Indoxyl sulfate (INDS) and p-cresol sulfate (pCS) are two of the most relevant uremic toxins that are recognized to have an essential role in chronic kidney disease (CKD) progression and associated cardiovascular risk. Thus, it is crucial to accurately assess their circulating levels in the body. Aiming at establishing an analytical strategy for quantification of INDS and pCS in human plasma, an automatic on-line micro-solid-phase extraction (μSPE) procedure hyphenated to tandem mass spectrometry (MS/MS) detection without previous chromatographic separation was herein developed. The bead injection (BI) concept was used to implement the μSPE procedure in the lab-on-valve (LOV) format. After studying the extraction conditions, the anion-exchange OASIS WAX sorbent beads (10 mg) and 99% ACN-H2O (15:85, v/v)-1% (v/v) NH4OH were chosen as sorbent and eluent, respectively, as they provided the highest analyte recoveries. Subsequently, the μSPE-BI-LOV system was hyphenated on-line to a MS/MS detector and the full analytical cycle, comprising sample preparation and analytes detection, was completed in <20 min. The developed μSPE-BI-LOV-MS methodology presented good linearity (r2 > 0.999) for quantification of the target analytes at concentrations ranging from 18 to 360 μg mL-1 in plasma. LOQ values were 2 μg mL-1 for INDS and 7 μg mL-1 for pCS in plasma. Human plasma samples from healthy subjects and individuals with CKD were successfully analyzed using the developed approach. The proposed automatic methodology can be described as an eco-friendly strategy, with a favorable score of 0.64 after greenness evaluation using the AGREE metric.

Evaluating 3D-printed bioseparation structures using multi-length scale tomography

X-ray computed tomography was applied in imaging 3D-printed gyroids used for bioseparation in order to visualize and characterize structures from the entire geometry down to individual nanopores. Methacrylate prints were fabricated with feature sizes of 500 µm, 300 µm, and 200 µm, with the material phase exhibiting a porous substructure in all cases. Two X-ray scanners achieved pixel sizes from 5 µm to 16 nm to produce digital representations of samples across multiple length scales as the basis for geometric analysis and flow simulation. At the gyroid scale, imaged samples were visually compared to the original computed-aided designs to analyze printing fidelity across all feature sizes. An individual 500 µm feature, part of the overall gyroid structure, was compared and overlaid between design and imaged volumes, identifying individual printed layers. Internal subvolumes of all feature sizes were segmented into material and void phases for permeable flow analysis. Small pieces of 3D-printed material were optimized for nanotomographic imaging at a pixel size of 63 nm, with all three gyroid samples exhibiting similar geometric characteristics when measured. An average porosity of 45% was obtained that was within the expected design range, and a tortuosity factor of 2.52 was measured. Applying a voidage network map enabled the size, location, and connectivity of pores to be identified, obtaining an average pore size of 793 nm. Using Avizo XLAB at a bulk diffusivity of 7.00 × 10-11 m2s-1 resulted in a simulated material diffusivity of 2.17 × 10-11 m2s-1 ± 0.16 × 10-11 m2s-1.

Scrutinizing different ionization responses of polar lipids in a reversed-phase gradient by implementing a counter-gradient

Lipidomics studies strive for a comprehensive identification and quantification of lipids. While reversed phase (RP) liquid chromatography (LC) coupled to high resolution mass spectrometry (MS) offers unrivalled selectivity and thus is the preferred method for lipid identification, accurate lipid quantification remains challenging. The widely adopted one-point lipid class specific quantification (one internal standard per lipid class) suffers from the fact that ionization of internal standard and target lipid occurs under different solvent composition as a consequence of chromatographic separation. To address this issue, we established a dual flow injection and chromatography setup that allows to control solvent conditions during ionization enabling isocratic ionization while running a RP gradient through the use of a counter-gradient. Using this dual LC pump platform, we investigated the impact of solvent conditions within a RP gradient on ionization response and arising quantification biases. Our results confirmed that changing solvent composition significantly influences ionization response. Quantification of human plasma (SRM 1950) lipids under gradient and isocratic ionization conditions further confirmed these findings as significant differences between the two conditions were found for the majority of lipids. While the quantity of sphingomyelins with >40 C atoms was consistently overestimated under gradient ionization, isocratic ionization improved their recovery compared to consensus values. However, the limitation of consensus values was demonstrated as overall only small changes in z-score were observed because of high uncertainties of the consensus values. Furthermore, we observed a trueness bias between gradient and isocratic ionization when quantifying a panel of lipid species standards which is highly dependent on lipid class and ionization mode. Uncertainty calculations under consideration of the trueness bias as RP gradient uncertainty revealed that especially ceramides with >40 C atoms had a high bias leading to total combined uncertainties of up to 54%. The assumption of isocratic ionization significantly decreases total measurement uncertainty and highlights the importance of studying the trueness bias introduced by a RP gradient to reduce quantification uncertainty.